Radeon drivers broken hrev46860

[vidrep]

Installed hrev46860 32 bit and 64 bit to hard drive partitions Intel Extreme driver working with both builds ATI Radeon HD 3470 DVI - 32 bit no signal

• 64 bit KDL on boot

ATI Radeon HD 3450 VGA - 32 bit no signal

• 64 bit no signal

Both video cards confirmed to work correctly on a4.1 using radeon drivers from dev build dated 2012/11/14.

[kallisti5]

"KERN: radeon_hd: connector_read_edid: invalid gpio 0 for connector 0"

I reworked the GPIO pins recently... looks like something got messed up somewhere. I'll take a look at this one asap.

[kallisti5]

*should* be fixed as of hrev46862. My card didn't have more than 16 GPIO pins so I didn't notice this issue. Thanks for reporting!

Give it a test and let us know if this one is resolved.

[vidrep]

I've attached the KDL from the 64 bit build

[vidrep]

I have some NVidia cards to test as well. If there is a problem I'll open a new ticket for those.

[vidrep]

Could this be somehow related to ticket #10535 - no warm boot?

[vidrep]

Similar issue with hrev46885 only now there is a white screen. See possible duplicate ticket #10563.

[kallisti5]

Please try again post-hrev46939

[vidrep]

hrev46947_gcc2_hybrid Radeon HD2400 - DMS-59 - screen flashing between Desktop and black (working on alpha4.1) Radeon HD3450 - DVI & VGA - working Radeon HD3550 - DVI - working

hrev46947_x86_64 Radeon drivers not working - black screen, no signal with all cards.

After booting and blue screen appears, the time it takes to see the Desktop is taking on average 10 seconds longer than alpha 4.1. Maybe an unrelated problem?

[vidrep]

The two Radeon cards described in the ticket (HD3450, HD3470) were working as of hrev46947. I'll open separate tickets for the non-working Radeon HD2400 and non-working 64 bit drivers described in last entry.

[kallisti5]

It seems like Julian Harnath's recent work broke radeon_hd across a large number of cards. I'm working on tracking down the offending commit to revert it.

[jua]

The commit in question would be hrev46988. The issues initially described in this bug ticket are older than my changes though, so these are definitely a separate problem.

I'm still quite sure that the change in hrev46988 is correct though. However it could be another bug that surfaced because of it. As said last night on IRC, it would be very useful to have a syslog of one of the machines where it broke because of hrev46988 so I could see what could be wrong.

[vidrep]

See ticket #10614 for my syslog using hrev46950.

[jua]

Ah, should have been more clear: the request for syslogs was to kallisti5 who said I broke it for many cards. Since your problem already occurs in hrev46950, it cannot have been introduced by my commit in hrev46988, that has to be something different.

[vidrep]

Perhaps this ticket should be modified to read "64 bit Radeon Drivers Broken", since the 32 bit drivers appear to have been fixed some time ago (see ticket #10563).

[diger]

hrev47098 x86_64, HD3000, DVI - black screen hrev47098 x86 - all ok

syslog

[vidrep]

The 32 bit Radeon drivers appear to work for some time now. It is the 64 bit drivers which continue to have a problem. This ticket should probably be closed and move the 64 bit driver issue to ticket #10614 which is 64 bit specific.